AS Modelling & Data Ltd. Prepared by Steve Smith. Telephone:

Transcription

1 A Dispersion Modelling Study of the Impact of Odour from the Proposed Poultry Houses at Land North of Nottingham Road, near Cropwell Bishop in Nottinghamshire Prepared by Steve Smith AS Modelling & Data Ltd. Telephone: th June

2 1. Introduction AS Modelling & Data Ltd. has been instructed by Ian Pick of Ian Pick Associates, on behalf of Highslade Property Ltd. to use computer modelling to assess the impact of odour emissions from the proposed broiler chicken rearing houses at Land North of Nottingham Road, near Cropwell Bishop in Nottinghamshire. NG12 2JU. Odour emission rates from the proposed poultry houses have been assessed and quantified based upon an emissions model that takes into account the likely internal odour concentrations and ventilation rates of the poultry houses. The odour emission rates so obtained have then been used as inputs to an atmospheric dispersion model which calculates odour exposure levels in the surrounding area. This report is arranged in the following manner: Section 2 provides relevant details of the site and potentially sensitive receptors in the area. Section 3 provides some general information on odour, details of the method used to estimate odour emissions from the proposed poultry houses, relevant guidelines and legislation on exposure limits and where relevant, details of likely background levels of odour. Section 4 provides some information about ADMS, the dispersion model used for this study and details the modelling parameters and procedures. Section 5 contains the results of the modelling. Section 6 provides a discussion of the results and conclusions. 2

3 2. Background Details The site of the proposed broiler rearing houses is adjacent to the A46, to the north of an Anaerobic Digestion Plant. The surrounding land is used primarily for arable farming, although there are some isolated wooded areas. The site is at an altitude of around 41 m; the land rises gently towards hill tops to the south and falls towards the Grantham Canal to the north-west. It is proposed that four new broiler chicken rearing houses be constructed at the site. These four new houses would provide accommodation for up to 220,000 broiler chickens and would be ventilated primarily by uncapped high speed ridge/roof mounted fans, each with a short chimney, with gable end fans which would provide supplementary ventilation in hot weather conditions. The chickens would be reared from day old chicks to up to around 38 days old and there would be approximately 7.5 flocks per annum. There are some isolated residences and commercial properties in the area surrounding the site of the proposed poultry houses. The closest are; a residence and garage, approximately 300 m to the southsouth-south-west and Quarry Farm, approximately 600 m to the south-east. A map of the surrounding area is provided in Figure 1; in the figure, the site of the proposed poultry houses is outlined in blue. 3

4 Figure 1. The area surrounding the site of the proposed poultry unit at land north of Nottingham Road Crown copyright and database rights

5 3. Odour, Emission Rates, Exposure Limits & Background Levels 3.1 Odour concentration, averaging times, percentiles and FIDOR Odour concentration is expressed in terms of European Odour Units per metre cubed of air (ou E/m 3 ). The following definitions and descriptions of how an odour might be perceived by a human with an average sense of smell may be useful, however, it should be noted that within a human population there is considerable variation in acuity of sense of smell. 1.0 ou E/m 3 is defined as the limit of detection in laboratory conditions. At ou E/m 3, particular odour might be detected against background odours in an open environment. When the concentration reaches around 5.0 ou E/m 3, a particular odour will usually be recognisable, if known, but would usually be described as faint. At 10.0 ou E/m 3, most would describe the intensity of the odour as moderate or strong and if persistent, it is likely that the odour would become intrusive. The character, or hedonic tone, of an odour is also important; typically, odours are grouped into three categories. Most offensive: Processes involving decaying animal or fish remains. Processes involving septic effluent or sludge. Biological landfill odours. Moderately offensive: Intensive livestock rearing. Fat frying (food processing). Sugar beet processing. Well aerated green waste composting. Less offensive: Brewery. Confectionery. Coffee roasting. Bakery. 5

6 Dispersion models usually calculate hourly mean odour concentrations and Environment Agency guidelines and findings from UK Water Industry Research (UKWIR) are also framed in terms of hourly mean odour concentration. The Environment Agency guidelines and findings from UKWIR use the 98 th percentile hourly mean; this is the hourly mean odour concentration that is equalled or exceeded for 2% of the time period considered, which is typically one year. The use of the 98 th percentile statistic allows for some consideration of both frequency and intensity of the odours. At some distance from a source, it would be unusual if odour concentration remained constant for an hour and in reality, due to air turbulence and changes in wind direction, short term fluctuations in concentration are observed. Therefore, although average exposure levels may be below the detection threshold, or a particular guideline, a population may be exposed to short term concentrations which are higher than the hourly average. It should be noted that a fluctuating odour is often more noticeable than a steady background odour at a low concentration. It is implicit that within the models hourly averaging time and the Environment Agency guidelines and findings from UKWIR that there would be variation in the odour concentration around this mean i.e. there would be short periods when odour concentration would be higher than the mean and lower than the mean. The FIDOR acronym is a useful reminder of the factors that will determine the degree of odour pollution: Frequency of detection. Intensity as perceived. Duration of exposure. Offensiveness. Receptor sensitivity. 3.2 Environment Agency guidelines In April 2011, the Environment Agency published H4 Odour Management guidance (H4). In Appendix 3 Modelling Odour Exposure, benchmark exposure levels are provided. The benchmarks are based on the 98 th percentile of hourly mean concentrations of odour modelled over a year at the site/installation boundary. The benchmarks are: 1.5 ou E/m 3 for most offensive odours. 3.0 ou E/m 3 for moderately offensive odours. 6.0 ou E/m 3 for less offensive odours. Any modelled results that project exposures above these benchmark levels, after taking uncertainty into account, indicates the likelihood of unacceptable odour pollution. 6

7 3.3 UK Water Industry Research findings The main source of research into odour impacts in the UK has been the wastewater industry. An indepth study of the correlation between modelled odour impacts and human response was published by UKWIR in This was based on a review of the correlation between reported odour complaints and modelled odour impacts in relation to nine wastewater treatment works in the UK with on-going odour complaints. The findings of this research and subsequent UKWIR research indicated the following, based on the modelled 98 th percentile of hourly mean concentrations of odour: At below 5.0 ou E/m 3, complaints are relatively rare at only 3% of the total registered. At between 5.0 ou E/m 3 and 10.0 ou E/m 3, a significant proportion of total registered complaints occur, 38% of the total. The majority of complaints occur in areas of modelled exposures of greater than 10.0 ou E/m 3, 59% of the total. 3.4 Choice of odour benchmarks for this study Odours from poultry rearing are usually placed in the moderately offensive category. Therefore, for this study, the Environment Agency s benchmark for moderately offensive odours, a 98 th percentile hourly mean of 3.0 ou E/m 3 over a one year period, is used to assess the impact of odour emissions from the proposed poultry unit at potentially sensitive receptors in the surrounding area. The UKWIR research is also considered. 3.5 Quantification of odour emissions Odour emission rates from broiler houses depend on many factors and are highly variable. At the beginning of a crop cycle, when chicks are small, litter is clean and only minimum ventilation is required, the odour emission rate may be small. Towards the end of the crop, odour production within the poultry housing increases rapidly and ventilation requirements are greater, particularly in hot weather, therefore emission rates are considerably greater than at the beginning of the crop. Peak odour emission rates are likely to occur when the housing is cleared of spent litter at the end of each crop. There is little available information on the magnitude of this peak emission, but it is likely to be greater than any emission that might occur when there are birds in the house. The time taken to perform the operation is usually around two hours per shed and it is normal to maintain ventilation during this time. There are measures that can be taken to minimise odour production whilst the housing is being cleared of spent litter and there is usually some discretion as to when the operation is carried out; therefore, to avoid high odour levels at nearby sensitive receptors it may be possible to time the operation to coincide with winds blowing in a favourable direction. To calculate an odour emission rate it is necessary to know the internal odour concentration and ventilation rate of the poultry house. For the calculation, the internal concentration is assumed to be a function of the age of the crop and the stocking density. 7

8 The internal concentrations used in the calculations increase exponentially from 300 ou E/m 3 at day 1 of the crop, to approximately 700 ou E/m 3 at day 16 of the crop, to approximately 1,800 ou E/m 3 at day 30 of the crop and approximately 2,300 ou E/m 3 at day 34 of the crop. These figures are obtained from a review of available literature and are based primarily on Robertson et al. (2002). The ventilation rates used in the calculations are based on industry practices and standard bird growth factors. Minimum ventilation rates are as those of an operational poultry house and maximum ventilation rates are based on Defra guidelines. Target internal temperature is 33 Celsius at the beginning of the crop and is decreased to 22 Celsius by day 34 of the crop. If the external temperature is 7 Celsius, or more, lower than the target temperature, minimum ventilation only is assumed for the calculation. Above this, ventilation rates are increased in proportion to the difference between ambient temperature and target internal temperature. A maximum transitional ventilation rate (35% of the maximum possible ventilation rate) is reached when the ambient temperature is equal to the target temperature. A high ventilation rate (70% maximum possible ventilation rate) is reached when the temperature is 4 degrees above target and if external temperature is above 33 Celsius the maximum ventilation rate is assumed. At high ventilation rates, it is likely that internal odour concentrations fall because odour is extracted much faster than it is created. Therefore, if the calculated ventilation rate exceeds that required to replace the volume of air in the house every 5 minutes, internal concentrations are reduced (by a factor of the square root of 7.5 times the shed volume/divided by the ventilation rate as an hourly figure). Based upon these principles, an emission rate for each hour of the period modelled is calculated by multiplying the concentration by the ventilation rate. Both the crop length and period the housing is empty can be varied. An estimation of the emission during the cleaning out process can also be included. In this case it is assumed that the houses are cleared sequentially and each house takes 2 hours to clear. In this case it is assumed for the calculations that the crop length is 38 days, with 30% thinning of the birds at day 32 and that there is an empty period of 11 days after each crop. To provide robust statistics, three sets of calculations were performed; the first with the first day of the meteorological record coinciding with day 1 of the crop cycle, the second coinciding with day 16 of the crop and the third coinciding with day 33 of the crop. A summary of the emission rates used in this study is provided in Table 1. It should be noted that the figures in this table refer to the whole of the crop length and that most figures quoted in literature are figures obtained from the latter stages of the crop cycle and therefore should not be compared directly to the AS Modelling & Data Ltd. figures in the table. The specific odour emission rate used for the clearing process is approximately 3.40 ou E/bird/s and the 98 th percentile emission rate is approximately 1.25 ou E/bird/s. As an example, a graph of the specific emission rate over the first year of the meteorological record for each of the three crop cycles is shown in Figure 2. 8

9 Table 1. Summary of odour emission rates (average/maximum of all 3 cycles) Emission rate (oue/s per bird as stocked, during crop) Season Average Night-time Average Day-time Average Maximum Winter Spring Summer Autumn Figure 2. Specific emission rate over the first year of each of the three crop cycles 9

10 4. The Atmospheric Dispersion Modelling System (ADMS) and Model Parameters The Atmospheric Dispersion Modelling System (ADMS) ADMS 5 is a new generation Gaussian plume air dispersion model, which means that the atmospheric boundary layer properties are characterised by two parameters; the boundary layer depth and the Monin-Obukhov length rather than in terms of the single parameter Pasquill-Gifford class. Dispersion under convective meteorological conditions uses a skewed Gaussian concentration distribution (shown by validation studies to be a better representation than a symmetrical Gaussian expression). ADMS has a number of model options including: dry and wet deposition; NO x chemistry; impacts of hills, variable roughness, buildings and coastlines; puffs; fluctuations; odours; radioactivity decay (and γ-ray dose); condensed plume visibility; time varying sources and inclusion of background concentrations. ADMS has an in-built meteorological pre-processor that allows flexible input of meteorological data both standard and more specialist. Hourly sequential and statistical data can be processed and all input and output meteorological variables are written to a file after processing. The user defines the pollutant, the averaging time (which may be an annual average or a shorter period), which percentiles and exceedance values to calculate, whether a rolling average is required or not and the output units. The output options are designed to be flexible to cater for the variety of air quality limits, which can vary from country to country and are subject to revision. 10

11 4.1 Meteorological data Computer modelling of dispersion requires hourly sequential meteorological data and to provide robust statistics the record should be of a suitable length; preferably four years or longer. The meteorological data used in this study is obtained from assimilation and short term forecast fields of the Numerical Weather Prediction (NWP) system known as the Global Forecast System (GFS). The GFS is a spectral model and data are archived at a horizontal resolution of 0.25 degrees, which is approximately 25 km over the UK (formerly 0.5 degrees, or approximately 50 km). The GFS resolution adequately captures major topographical features and the broad-scale characteristics of the weather over the UK. Smaller scale topological features may be included in the dispersion modelling by using the flow field module of ADMS (FLOWSTAR). The use of NWP data has advantages over traditional meteorological records because: Calm periods in traditional records may be over represented, this is because the instrumentation used may not record wind speeds below approximately 0.5 m/s and start up wind speeds may be greater than 1.0 m/s. In NWP data, the wind speed is continuous down to 0.0 m/s, allowing the calms module of ADMS to function correctly. Traditional records may include very local deviations from the broad-scale wind flow that would not necessarily be representative of the site being modelled; these deviations are difficult to identify and remove from a meteorological record. Conversely, local effects at the site being modelled are relatively easy to impose on the broad-scale flow and provided horizontal resolution is not too great, the meteorological records from NWP data may be expected to represent well the broad-scale flow. Information on the state of the atmosphere above ground level which would otherwise be estimated by the meteorological pre-processor may be included explicitly. The wind rose for the raw GFS data at the site of the proposed poultry unit is shown in Figure 3a. Wind speeds are modified by the treatment of roughness lengths (see Section 4.7) and because terrain data is included in the modelling the raw GFS wind speeds and directions will be modified. The terrain and roughness length modified wind rose for the site of the poultry unit is shown in Figure 3b. Note that, although there is very little modification in this case, elsewhere in the modelling domain modified wind roses may differ more markedly and that the resolution of the wind field in terrain runs is 200 m. 11

12 Figure 3a. The wind rose. Raw GFS derived data for N, W, (knots) Wind speed (m/s) 12

13 Figure 3b. The wind rose. FLOWSTAR modified GFS derived data for NGR , CORE-I5\Work\Projects\Fields_Farm\modelling\FLOWSTAR_52.916_-1.002_ADMS_010113_ (knots) Wind speed (m/s) 13

14 4.2 Emission sources Emissions from the chimneys of uncapped high speed ridge/roof fans on the proposed poultry houses are represented by three point sources per house within ADMS (PR1 a, b & c to PR4 a, b & c). Details of the point source parameters are shown in Table 2a. The positions of the point sources may be seen in Figure 4. Table 2. Point source parameters Source ID Height (m) Diameter (m) Efflux velocity (m/s) Emission temperature (ºC) Emission rate per source (ou E/s) PR1 a, b & c to PR4 a, b & c Variable 1 Variable 1 1. Dependent on ambient temperature. 2. Reduced by 50% when the ambient temperature equals or exceeds 21 Celsius. The proposed poultry houses are would also be fitted with gable end fans which would be used to provide supplementary ventilation in hot weather conditions. The emissions from these gable end fans are represented by a single volume source within ADMS (GABLES). Details of the volume source parameters are shown in Table 2b. The position of the volume sources may be seen in Figure 4. Table 2b. Volume source parameters Source ID (Scenario) Length Y (m) Width X (m) Depth (m) Base height (m) Emission temperature ( C) Emission rate (ou E/s) GABLES Ambient Variable % of the total emission is emitted when the ambient temperature equals or exceeds 21 Celsius. 4.3 Modelled buildings The structure of the proposed poultry houses and nearby buildings and tanks at the AD plant may affect the odour plumes from the point sources. Therefore, the buildings are modelled within ADMS. The positions of the modelled buildings may be seen in Figure 4 where they are marked by grey rectangles. 4.4 Discrete receptors Sixteen discrete receptors have been defined at a selection of nearby residences and commercial properties that are within approximately 1,000 m of the site of the proposed poultry houses. The receptors are defined at 1.5 m above ground level within ADMS and their positions may be seen in Figure 5 where they are marked by enumerated pink rectangles. 4.5 Nested Cartesian grid To produce the contour plots presented in Section 5 of this report, a nested Cartesian grid has been defined within ADMS. The grid receptors are defined at 1.5 m above ground level within ADMS. The positions of the receptors may be seen in Figure 5 where they are marked by green crosses. 14

15 4.6 Terrain data There are some slopes and hills that would affect wind flow and dispersion of odour in the area around the site of the proposed poultry unit; therefore, terrain has been considered in the modelling. The terrain data used are derived from the Ordnance Survey 50 m Digital Elevation Model. These data are resampled at a 50 m horizontal resolution for use within ADMS. The terrain domain is 6.4 km by 6.4 km and FLOWSTAR is run at a resolution of 32 x 32 points; therefore, the effective model resolution is 200 m. 4.7 Other model parameters A fixed surface roughness length of 0.25 m has been applied over the entire modelling domain. As a precautionary measure, the GFS meteorological data is assumed to have a roughness length of m. The effect of the difference in roughness length is precautionary as it increases the frequency of low wind speeds and the stability and therefore increases predicted ground level concentrations. 15

16 Figure 4. The positions of modelled buildings & sources Crown copyright and database rights

17 Figure 5. The discrete receptors and nested Cartesian grid receptors Crown copyright and database rights

18 5. Details of the Model Runs and Results For this study ADMS was run with the terrain module of ADMS (FLOWSTAR) and with the calms module of ADMS. ADMS was effectively run twelve times, once for each year of the four year meteorological record and for each of the broiler rearing cycles. Statistics for the annual 98 th percentile hourly mean odour concentration at each receptor were compiled for each of the twelve runs. A summary of the results of these twelve runs at the discrete receptors is provided in Table 3a where the maximum annual 98 th percentile hourly mean odour concentration is shown. A contour plot of the maximum annual 98 th percentile hourly mean odour concentration is shown in Figure 6. In Table 3, predicted odour exposures in excess of the Environment Agency s benchmark of 3.0 ou E/m 3 as an annual 98 th percentile hourly mean are coloured blue; those in the range that UKWIR research suggests gives rise to a significant proportion of complaints, 5.0 ou E/m 3 to 10.0 ou E/m 3 as an annual 98 th percentile hourly mean, are coloured orange and predicted exposures likely to cause annoyance and complaint are coloured red. Table 3. Predicted maximum annual 98 th percentile hourly mean odour concentrations at the discrete receptors Receptor number X(m) Y(m) Maximum annual 98 th percentile hourly mean odour concentration (oue/m 3 ) GFS Calms Terrain

19 Figure 6. Predicted maximum annual 98 th percentile hourly mean odour concentration Crown copyright and database rights

20 6. Summary and Conclusions AS Modelling & Data Ltd. has been instructed by Ian Pick of Ian Pick Associates, on behalf of Highslade Property Ltd. to use computer modelling to assess the impact of odour emissions from the proposed broiler chicken rearing houses at Land North of Nottingham Road, near Cropwell Bishop in Nottinghamshire. NG12 2JU. Odour emission rates from the proposed poultry houses have been assessed and quantified based upon an emissions model that takes into account the likely internal odour concentrations and ventilation rates of the poultry houses. The odour emission rates so obtained have then been used as inputs to an atmospheric dispersion model which calculates odour exposure levels in the surrounding area. The results of the modelling indicate that, should the proposed development of the poultry proceed, the 98 th percentile hourly mean odour concentration at nearby residences would be below the Environment Agency s benchmark for moderately offensive odours, a 98 th percentile hourly mean of 3.0 ou E/m 3 over a one year period. 20

21 7. References Environment Agency, April H4 Odour Management, How to comply with your environmental permit. Chartered Institution of Water and Environmental Management website. Control of Odour. R. E. Lacey, S. Mukhtar, J. B. Carey and J. L. Ullman, A Review of Literature Concerning Odors, Ammonia, and Dust from Broiler Production Facilities. M. Navaratnasamy. Odour Emissions from Poultry Manure/Litter and Barns. Fardausur Rahaman et al. ESTIMATION OF ODOUR EMISSIONS FROM BROILER FARMS AN ALTERNATIVE APPROACH A. P. Robertson et al, Commercial-scale Studies of the Effect of Broiler-protein Intake on Aerial Pollutant Emissions. ROSS. Environmental Management in the Broiler House Defra. Heat Stress in Poultry - Solving the Problem 21